swift-aws-lambda-runtime/Examples/JSONLogging

JSON Logging Example

This example demonstrates how to use structured JSON logging with AWS Lambda functions written in Swift. When configured with JSON log format, your logs are automatically structured as JSON objects, making them easier to search, filter, and analyze in CloudWatch Logs.

Features

• Structured JSON log output
• Automatic inclusion of request ID and trace ID
• Support for all log levels (TRACE, DEBUG, INFO, WARN, ERROR, FATAL)
• Custom metadata in logs
• Compatible with CloudWatch Logs Insights queries

Code

The Lambda function demonstrates various logging levels and metadata usage. When AWS_LAMBDA_LOG_FORMAT is set to JSON, all logs are automatically formatted as JSON objects with the following structure:

{
  "timestamp": "2024-10-27T19:17:45.586Z",
  "level": "INFO",
  "message": "Processing request for Alice",
  "requestId": "79b4f56e-95b1-4643-9700-2807f4e68189",
  "traceId": "Root=1-67890abc-def12345678901234567890a"
}

Configuration

Environment Variables

• AWS_LAMBDA_LOG_FORMAT: Set to JSON for structured logging (default: Text)
• AWS_LAMBDA_LOG_LEVEL: Control which logs are sent to CloudWatch
  • Valid values: TRACE, DEBUG, INFO, WARN, ERROR, FATAL
  • Default: INFO when JSON format is enabled

SAM Template Configuration

Add the LoggingConfig property to your Lambda function:

Resources:
  JSONLoggingFunction:
    Type: AWS::Serverless::Function
    Properties:
      CodeUri: .build/plugins/AWSLambdaPackager/outputs/AWSLambdaPackager/JSONLogging/JSONLogging.zip
      Handler: swift.bootstrap
      Runtime: provided.al2023
      Architectures:
        - arm64
      LoggingConfig:
        LogFormat: JSON
        ApplicationLogLevel: INFO  # TRACE | DEBUG | INFO | WARN | ERROR | FATAL
        SystemLogLevel: INFO       # DEBUG | INFO | WARN

Test Locally

Start the local server with TEXT logging:

swift run

Send test requests:

# Basic request
curl -d '{"name":"Alice"}' http://127.0.0.1:7000/invoke

# Request with custom level
curl -d '{"name":"Bob","level":"debug"}' http://127.0.0.1:7000/invoke

# Trigger error logging
curl -d '{"name":"error"}' http://127.0.0.1:7000/invoke

To test with JSON logging locally, set the environment variable:

AWS_LAMBDA_LOG_FORMAT=JSON swift run

Build & Package

swift build
LAMBDA_USE_LOCAL_DEPS=../.. swift package archive --allow-network-connections docker

The deployment package will be at: .build/plugins/AWSLambdaPackager/outputs/AWSLambdaPackager/JSONLogging/JSONLogging.zip

Deploy with SAM

Create a template.yaml file:

AWSTemplateFormatVersion: '2010-09-09'
Transform: AWS::Serverless-2016-10-31
Description: JSON Logging Example

Resources:
  JSONLoggingFunction:
    Type: AWS::Serverless::Function
    Properties:
      CodeUri: .build/plugins/AWSLambdaPackager/outputs/AWSLambdaPackager/JSONLogging/JSONLogging.zip
      Timeout: 60
      Handler: swift.bootstrap
      Runtime: provided.al2023
      Architectures:
        - arm64
      LoggingConfig:
        LogFormat: JSON
        ApplicationLogLevel: DEBUG
        SystemLogLevel: INFO

Outputs:
  FunctionName:
    Description: Lambda Function Name
    Value: !Ref JSONLoggingFunction

Deploy:

sam deploy --guided

Deploy with AWS CLI

As an alternative to SAM, you can use the AWS CLI:

ACCOUNT_ID=$(aws sts get-caller-identity --query 'Account' --output text)
aws lambda create-function \
  --function-name JSONLoggingExample \
  --zip-file fileb://.build/plugins/AWSLambdaPackager/outputs/AWSLambdaPackager/JSONLogging/JSONLogging.zip \
  --runtime provided.al2023 \
  --handler swift.bootstrap \
  --architectures arm64 \
  --role arn:aws:iam::${ACCOUNT_ID}:role/lambda_basic_execution \
  --logging-config LogFormat=JSON,ApplicationLogLevel=DEBUG,SystemLogLevel=INFO

Invoke

aws lambda invoke \
  --function-name JSONLoggingExample \
  --cli-binary-format raw-in-base64-out \
  --payload '{"name":"Alice","level":"debug"}' \
  response.json && cat response.json && rm response.json

Query Logs with CloudWatch Logs Insights

With JSON formatted logs, you can use powerful queries in CloudWatch Logs Insights.

Using the AWS Console

1. Open the CloudWatch Logs Insights console
2. In the "Select log group(s)" dropdown, choose the log group for your Lambda function (typically /aws/lambda/JSONLoggingExample)
3. Type or paste one of the queries below into the query editor
4. Adjust the time range in the top-right corner to cover the period you're interested in
5. Click "Run query"

# Find all ERROR level logs
fields @timestamp, level, message, requestId
| filter level = "ERROR"
| sort @timestamp desc

# Find logs for a specific request
fields @timestamp, level, message
| filter requestId = "79b4f56e-95b1-4643-9700-2807f4e68189"
| sort @timestamp asc

# Count logs by level
stats count() by level

# Find logs with specific metadata
fields @timestamp, message, metadata.errorType
| filter metadata.errorType = "SimulatedError"

Using the AWS CLI

You can also run Logs Insights queries from the command line. Each query is a two-step process: start the query, then fetch the results.

# 1. Start a query (adjust --start-time and --end-time as needed)
QUERY_ID=$(aws logs start-query \
  --log-group-name '/aws/lambda/JSONLoggingExample' \
  --start-time $(date -v-1H +%s) \
  --end-time $(date +%s) \
  --query-string 'fields @timestamp, level, message | filter level = "ERROR" | sort @timestamp desc' \
  --query 'queryId' --output text)

# 2. Wait a moment for the query to complete, then get the results
sleep 2
aws logs get-query-results --query-id "$QUERY_ID"

A few more examples:

# Count logs by level over the last 24 hours
QUERY_ID=$(aws logs start-query \
  --log-group-name '/aws/lambda/JSONLoggingExample' \
  --start-time $(date -v-24H +%s) \
  --end-time $(date +%s) \
  --query-string 'stats count() by level' \
  --query 'queryId' --output text)
sleep 2
aws logs get-query-results --query-id "$QUERY_ID"

# Find logs with a specific error type in the last hour
QUERY_ID=$(aws logs start-query \
  --log-group-name '/aws/lambda/JSONLoggingExample' \
  --start-time $(date -v-1H +%s) \
  --end-time $(date +%s) \
  --query-string 'fields @timestamp, message, metadata.errorType | filter metadata.errorType = "SimulatedError"' \
  --query 'queryId' --output text)
sleep 2
aws logs get-query-results --query-id "$QUERY_ID"

Note

: On Linux, replace date -v-1H +%s with date -d '1 hour ago' +%s (and similarly for other time offsets).

Log Levels

The runtime maps Swift's Logger.Level to AWS Lambda log levels:

Swift Logger.Level	JSON Output	Description
.trace	TRACE	Most detailed
.debug	DEBUG	Debug information
.info	INFO	Informational
.notice	INFO	Notable events
.warning	WARN	Warning conditions
.error	ERROR	Error conditions
.critical	FATAL	Critical failures

Benefits of JSON Logging

1. Structured Data: Logs are key-value pairs, not plain text
2. Easy Filtering: Query specific fields in CloudWatch Logs Insights
3. Automatic Context: Request ID and trace ID included automatically
4. Metadata Support: Add custom fields to logs
5. No Double Encoding: Already-JSON logs aren't double-encoded
6. Better Analysis: Automated log analysis and alerting

Clean Up

# SAM deployment
sam delete

# AWS CLI deployment
aws lambda delete-function --function-name JSONLoggingExample

⚠️ Important Notes

• JSON logging adds metadata, which increases log size
• Default log level is INFO when JSON format is enabled
• For Python functions, the default changes from WARN to INFO with JSON format
• Logs are only formatted as JSON in the Lambda environment, not in local testing (unless you set AWS_LAMBDA_LOG_FORMAT=JSON)